3 ms·
Inspired by a Kaggle competition (https://www.kaggle.com/c/sf-crime https://www.kaggle.com/c/sf-crime), one of my older blog posts involved predicting the type
by minimaxir 9y ago
Inspired by a Kaggle competition (https://www.kaggle.com/c/sf-crime https://www.kaggle.com/c/sf-crime), one of my older blog posts involved predicting the type of arrest in San Francisco (given that an arrest occurred) using data such as location and timing and the relatively new LightGBM machine learning algorithm: http://minimaxir.com/2017/02/predicting-arrests/ http://minimaxir.com/2017/02/predicting-arrests/
The code is open-sourced in an R Notebook: http://minimaxir.com/notebooks/predicting-arrests/ http://minimaxir.com/notebooks/predicting-arrests/
The model performance isn't great enough to usher in precrime, even in the best case. There are likely better approaches nowadays. (e.g. since the location data is spatial, a convolutional neural network might work better.)